Skip to content

[TAS] Investigate LeastFreeCapacity algorithm for TAS podset groups #8178

@PBundyra

Description

@PBundyra

What happened:
During static analysis, I've encountered piece of code that I think may lead to undesired behavior of LeastFreeCapacity algorithm when using podset groups. When using LeastFreeCapacity In sortedDomainsWithLeader function we don't sort ascending by the stateWithLeader field which I believe we should:

https://github.com/kubernetes-sigs/kueue/blob/main/pkg/cache/scheduler/tas_flavor_snapshot.go#L1308-L1319

This need confirmation and then unit tests to ensure it wai.

What you expected to happen:

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?:

Environment:

  • Kubernetes version (use kubectl version):
  • Kueue version (use git describe --tags --dirty --always):
  • Cloud provider or hardware configuration:
  • OS (e.g: cat /etc/os-release):
  • Kernel (e.g. uname -a):
  • Install tools:
  • Others:

Metadata

Metadata

Assignees

Labels

kind/bugCategorizes issue or PR as related to a bug.

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions